For many years it has been apparent that transitional cell carcinomas are a heterogeneous group of neoplasms with two clinical forms that exhibit distinctly different prognoses. Approximately 20% of the tumors are invasive at presentation, which is associated with poor prognoses. The remaining carcinomas are superficial, and an excellent outcome can be expected in the majority of patients treated with local therapies. However, 20% of the latter will progress to muscle invasive disease during the follow-up. The problem is to identify those who will progress and to distinguish those tumors likely to respond to therapy. Genetic changes that identify the subgroups of these tumors may be the key issue. During the past decade, studies of human cancer have begun to yield molecular information on the identity of multiple genetic changes that underline development and progression. Attention was focused initially on oncogenes and more recently on tumor-suppressor genes.